Geochemical models suggest that the calcium carbonate saturation state (?CaCO3) and total alkalinity (TA) of the ocean fluctuated considerably in the geologic past, under the influence of the continental weathering HCO3- source and the biological CaCO3 sink. The mid-Mesozoic is a key period in this respect because the advent of planktic calcifying organisms introduced the calcium carbonate compensation mechanism, buffering oceanic ?CaCO3 and TA with respect to perturbations of the carbonate cycle. As a result, ?CaCO3 and TA are thought to have fallen drastically in the mid-Mesozoic, and, excluding short-lived perturbations of the carbon cycle, to have remained similar to present values ever since. This important idea, put forward using geochemical models, is tested here using the geological record of marine calcified cyanobacteria. It is shown that under the most conservative of assumptions oceanic ?CaCO3 must have been ? 6-10, and TA ? 3 mM, in the Permo-Triassic, prior to the advent of planktic calcifying organisms. It is likely, though, that oceanic ?CaCO3 (? 10-16) and TA (? 4-6 mM) were greater than these minimum estimates during most of this time interval. In the Late Cretaceous, once the pelagic CaCO3 sink was well in place, oceanic ?CaCO3 (~ 4) and TA (~ 2mM) had fallen to values compatible with those of the modern ocean. This analysis implies a large fall in oceanic ?CaCO3 and TA between 180 and 90 Myr, a time interval that encompasses the advent of planktic calcifying organisms and the introduction of the carbonate compensation mechanism. This cyanobacteria-based reconstruction strengthens the claim that pelagic calcification revolutionised the carbonate cycle. It also highlights, without implying a cause-effect link, that biocalcifying organisms evolved in an ocean with an elevated HCO3-/H+ ratio, a condition known to favour biological calcification in laboratory experiments.